基于USB650南疆红枣颜色分级方法的研究  

Research of Southern Xinjiang Jujube Color Classification Methods Based on the USB650

作　　者：彭云发 解海龙[1] 罗华平[1,2] 

机构地区：[1]塔里木大学机械电气化工程学院,新疆阿拉尔843300 [2]新疆维吾尔自治区普通高等学校现代农业工程重点实验室,新疆阿拉尔843300

出　　处：《塔里木大学学报》2014年第4期81-85,共5页Journal of Tarim University

基　　金：国家自然基金项目资助(10964009;11164023)

摘　　要：以南疆红枣颜色快速分级为研究对象，采用美国海洋光学USB650红潮光谱仪来测量红枣的颜色特征值，采用红枣颜色按L[微软拥护1]值分级时，一级红枣的L值为30．2—35．4，二级红枣的L值为35．5～39．9，三级红枣的L值为40．0～44．6，按L值分级与人工分级的一致度为93．65％；以红枣的L、a、b值作为特征值，应用BP人工神经网络进行分级，BP人工神经网络分级与人工分级的一致度达到了94．76％。结果表明，以L、a、b三个颜色值分级效果要好于仅以L值分级，BP人工神经网络分级效果优于人工分级，较好地满足红枣颜色分级的需求，对南疆红枣产品的加工、销售具有一定的理论指导和实际应用意义。With jujube color rapid classification as the research object, using the United States ocean optics USB650 red tide spectrometer to measure the color characteristic values of jujube. The jujube color classification according to the L value, the L value of first grade jujube is 30. 2 -35. 4, the second grade jujube is 35. 5 -39. 9 ,the third grade jujube is 40. 0 -44. 6. The consistency according to the L value and the artificial classification is 93. 65%. Furthermore, putting the L, a, b value of jujube as the characteristic value, application of BP artificial neural network for classification, the consistency according to the BP artificial neural network and the artificial classification is 94. 76%. As the results shown, the classification effect with L, a and b three color value is better than only in the L value classification, the BP artificial neural network classification effect is good, and can meet the needs of jujube color grading better, having a certain theoretical guidance and actual application significance for southern Xinjiang jujube production and sales.

关 键 词：南疆红枣 颜色分级 神经网络 CIE 1976 L^＊a^＊b^＊颜色空间 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]